← Back to blog
Blog

When does the second order come — use your own orders to pick the day you follow up

If you send your follow-up email or coupon a set number of days after the first purchase because another store does it that way, you are probably guessing. When the second order arrives depends on what you sell. Count the days from first to second order in your own data, send one message just before the peak, and judge it by comparing groups of customers who first bought in different months.

7 min read
  • Reading the numbers

This continues the third of the five numbers to check each week: average order value. That post said that if AOV drops while the share of new customers rises, the thing to work on isn’t AOV — it’s how to get those new customers to buy a second time. This one is the detail: how many days after the first purchase to send the follow-up that asks for the second.

The usual advice is a fixed rule: email two weeks after purchase, send a coupon after a month. Easy to apply, but when a second order arrives depends heavily on what you sell. Borrow another store’s number and your message lands either before the customer needs anything, or after they’ve already reordered somewhere else.

What decides it is how many days your own customers take to go from first order to second.

1. The second order arrives at a different time in every store

Take two made-up stores. In each, 200 people placed their first order in a given month. Here is when each of those 200 placed a second order over the following 90 days, counted in days since the first.

Days from first to second orderConsumables store (coffee beans)Gift and homeware store
0–15 days6 people8 people
16–30 days22 people7 people
31–45 days30 people6 people
46–60 days10 people7 people
61–90 days8 people12 people
No second order within 90 days124 people160 people
Total200 people200 people

At the coffee store, second orders bunch up in the 31–45 day band — about when the first bag runs out. If this store sends a coupon two weeks after purchase, it arrives while there are still beans in the cupboard. There’s no reason to buy yet, and the coupon may simply expire.

The gift store has no clear peak. Second orders trickle in over several months. Fine-tuning the send day here may not buy much. The question to ask first isn’t when — it’s how many customers come back at all.

In the free-shipping threshold post, the point was that average order value doesn’t tell you where orders pile up. Same here: boil it down to “customers return after X days on average” and you can no longer see whether there’s a peak at all.

Check: before choosing a follow-up day, see which of these two shapes your own second orders take.

2. Count the gap between first and second order in half-month bands

There is one thing to do. Export your order list from your store’s admin and, for each customer, count the days between their first order and their second.

  • Include everyone who placed a first order in a given month. Combining two or three months is fine
  • Use half-month (15-day) bands. Single days leave too few orders per band and the shape jumps around
  • Keep customers who haven’t placed a second order in the denominator. Counting only people who came back makes the return rate look far higher than it is, and you’ll misread how tall the peak is
  • Customers whose first order was less than 90 days ago aren’t finished yet. Only count monthly groups that are at least 90 days past their first order

Also check that one customer is counted as one person. Stores with lots of guest checkouts often record the same shopper as several customers. Grouping by email address can turn up second orders you didn’t know about.

Check: list the number of people in each band plus those with no second order, and make sure they add up to the number of first-time buyers.

3. If there’s a peak, send one message just before it

Once the bands show a peak, the first candidate is one message sent shortly before the peak begins. For the coffee store in section 1, the peak is the 31–45 day band, so a send around day 25 reaches people as their beans are running low. Sending in the middle of the peak or after it mostly reaches people who’ve already reordered — from you or from someone else.

Send just one. Stack up a second and third message and, when you check the results in section 4, you won’t be able to tell which one did the work.

If there’s no peak, before tuning the day, look at the share of first-time buyers who place a second order within 60 days. From the table in section 1, that’s 68 out of 200 at the coffee store and 28 out of 200 at the gift store. A low share at the gift store isn’t necessarily about timing. The store may simply have few reasons for a second purchase — things that get used up, topped up, or swapped with the seasons. In that case, changing what the follow-up offers, from a coupon to something that goes with the first purchase, may do more than moving the send day.

Check: with a peak, pick one day just before it. Without one, write down your 60-day second-order rate first.

4. Judge it by comparing groups who first bought in different months

Once you’ve started sending, you need to know whether it worked. Looking at sales in the four weeks after you started will mislead you: those weeks also include repeat orders from people who first bought before you started, plus whatever the season is doing.

Compare groups of customers by the month of their first order instead.

  • People who first bought in the month before you started sending (say, the September group)
  • People who first bought in the month you started (say, the October group)

For each group, compare the share who placed a second order within 60 days of their first. The October group’s answer isn’t in until late December, when the last October buyers reach day 60. That feels like a long wait, but without comparing groups you can’t separate “the follow-up worked” from “it happened to be a busy month.” The traffic source post made the point that a total moves just because the mix inside it changes. Here, too, look at each group’s rate, not total sales.

If the rate went up, the timing is right. If it didn’t move, shift the send day a little closer to the start of the peak, or stop sending. If the follow-up is a coupon, check the value of those second orders as well — a higher rate can still mean you discounted orders that would have happened anyway.

Check: put the 60-day second-order rate for the group before you started next to the group after.

5. GA4’s new vs returning split can’t give you this gap

Every decision above comes from each customer’s order dates. GA4 does have a “new vs returning” split, but it decides whether two visits are the same person by browser cookie. In Safari, cookies set by scripts expire after 7 days, so a customer who comes back after more than a week is counted as new. According to statcounter, as of August 2026 Safari had 48.7% of mobile browser share in Japan. For a store like the coffee store, where second orders come 31–45 days later, many of those second orders show up in GA4 as first purchases.

So what you’re missing in order to pick a follow-up day is the first and second order dates for the same customer. Without that, you can’t find the peak, and once you’ve started sending you can’t tell whether it worked. You need your store’s order data.

When order data is connected, NextRise Analytics builds revenue and AOV from your store’s orders. Start with the Sales tab in the sample store and look at AOV and revenue by product.

It’s also worth checking whether second orders cluster just under your free-shipping threshold — that makes it easier to choose the one item to suggest in the follow-up. How to count the orders sitting just below the line is in Don’t set your free-shipping threshold just above average order value.

Related articles